Stop 1 of 9 · Why can't position and transport both be sharp on a closed axis?

Position and transport on a closed axis — every torus has its star

On the three-torus every axis is closed: the last position is linked to the first. Each axis carries a position P and a transport T°, and the two do not commute, so no state makes both sharp. Robertson's 1934 inequality says how much: the determinant of the fluctuations is at least the determinant of the brackets. With three positions per axis, at the maximal current state the ratio is 125/64. Pick a torus, run the transport, and drag either panel to turn both.

The torus: Ψ on every position

size |Ψ|² · colour phase

The star of the torus

δT° and δP of each axis

The angle at c on every ring

At the maximal current state every closed axis opens an angle θ, never zero: the dark share is positive. With 3 positions it is exactly arccos(2/√5) = 26.57° (light 4/5, Lean); with 4 or more positions it is computed here, between about 45° and 50°: the browser finds c as the top eigenvector of K = i[T°, P] (Jacobi diagonalization, double precision) and reads the angle from it. ○ θ at c · ● θ at Ψ(t). The page runs on its own: the transport bar fills once, then a new random torus, and the bar fills again. Press Tour to stop on a torus, Transport to pause.

The octahedron of δ

drag to turn · dashed: the largest δ

Each closed axis with d positions carries δ(d) = 1/cos θ − 1 at c. Put a vertex at ±δ on every axis: the octahedron of the torus. The canonical 3 × 3 × 3 gives the smallest, δ(3) = √5/2 − 1 ≈ 0.11803 on every axis (exact). Its volume is (4/3)·𝒱, with 𝒱 = δ(dx) δ(dy) δ(dz); the values for 4 or more positions are computed here.

δ on x, y, z–
𝒱 = δxδyδz–
𝒱 / δ(3)³ (T³ = 1)–
𝒱 / (largest δ)³–

Tension and the angle

|t(t)| against θ(t), one panel per axis

The tension of an axis is t = ⟨K⟩, K = i[T°, P]. In the band |t| > 1 every state keeps a positive dark share (TorusDark.darkShare_pos_of_band) and det|RS^N³| is strict on T³ for every state with the three axes in the band (TorusStrict.robertson_det_torus_lt). The maximal tension is the state c: √6/2 with 3 positions (ClosedThree.tension_maxState); for every d it lies in [1, 1 + 2/(d − 1)] and tends to 1 as the ring grows (ClosedAxis.one_le_cMax, ClosedAxis.cMax_le, ClosedAxis.tendsto_cMax): the current concentrates on the seam. Press Transport: Ψ(t) starts at c and leaves it.

family–
orbit under axis permutations–
dark share at c on x, y, z–
the dark share–
axispositionsθ(t)θ at c 1/cos θ at cσTσP(t)⟨P⟩(t)

Position by position: where does the probability go?

Ask a yes/no question: "is the excitation at position j?" Its chance is p_j (the violet disc). When the transport runs, that chance can only change by flowing through the two links of j (the orange arrows), like water between neighbouring tanks: what one position loses, its neighbour gains, and the total never changes. And there is a speed limit for each question: a position that is surely full, or surely empty, is not changing at that moment (the dashed frames). Press Transport; then try a state of definite energy.

What is exact here

Theorem trail
StatementName
Robertson 1934 on the torus, every stateTorus.robertson_det_torus
strict with every axis in its bandTorusStrict.robertson_det_torus_lt
on every axis, tension above 1 forces a positive defectTorus.gramDefect_pos
(5/4)^k at the maximal current state; 125/64; the ladderTorus.det_covMatrix_maxState, Torus.three_torus, Torus.ladder
different axes commuteTorusCommute.torusPairs_comm
continuity on a closed axis; nothing is lost; velocity = sum of flowsEventsTopology.rate_siteEvent, EventsTopology.sum_rate_siteEvent, EventsTopology.velocity_eq_sum_flow
Robertson for each position; a decided position does not changeEventsTopology.flow_sq_le, SiteEvents.rate_site_sq_le, SiteEvents.rate_eq_zero_of_decided
the four topologiesSiteEventsTopologies.rate_openCube, SiteEventsTopologies.rate_torus, SiteEventsTopologies.rate_box, TorusFiberDirac.rate_torusFiber

The pathStop 2 →